From scattered field notes, photographs of excavations in progress, and published and unpublished reports, it is possible to develop a composite picture of the field approach used during the Somerset County relief excavations. This field methodology varied from site to site and even during the excavation of individual sites. Nonetheless, extant field notes clearly indicate some attempt was made to follow relatively standard field excavation and recording techniques. This section outlines Somerset County relief excavation field methodology, including the selection of locations for excavation, field recording techniques, and excavation procedures.

    One goal of the Somerset County relief excavations was to disprove the then prevailing notion that there had never been a substantial prehistoric occupation by Native Americans in Somerset County (Butler 1936:55). Therefore, locations selected for excavation emphasized potential village sites, though a camp site and a number of rockshelters were also excavated. The emphasis on selecting Late Prehistoric Monongahela (ca. A.D. 900 to 1630s) village sites for excavation was related to at least two factors. First, Monongahela villages represented instances where a large number of people (at least compared to earlier periods) congregated into select locations and constructed several contiguous and contemporary structures. Consequently, Monongahela villages created a greater impact on their local environment than was seen in earlier periods. This impact made Monongahela villages more visible from  an archaeological perspective. Second, the structure of Monongahela villages ensured that a large number of features and relatively high quantity of artifacts would be encountered during excavation. A typical Somerset Monongahela community pattern consisted of a habitation zone containing activity areas, houses, pits, burials, and other features surrounding an open plaza area (Johnson et al., 1989:9). Thus, given the nature of Monongahela village sites, one was ensured that excavating a village site would keep a field crew occupied for some time. While advancing archaeological knowledge was the stated purpose of the Somerset County relief excavations, keeping people employed was an equally important, if not more important, goal. If the Somerset County field projects had not shown that they could keep a crew employed for a reasonable duration, it surely would have occurred to federal overseers that funding these archaeological investigations was not helping alleviate the unemployment problem.

    Potential village locations were selected from informant interviews collected during the 1934 CWA "paper" survey of Somerset County. These locations were then verified in the field. It is no accident that all excavated village sites were in cultivated fields. Archaeologists of the time understood that a correlation existed between artifacts on the surface of a plowed field and the type of cultural remains below the surface. This is clear in Augustine's (1939b:5) comments on the discovery of Fort Hill (36 SO 2):

Following up and down in the potato rows we noted small bits of pottery, flint chips and  bits of sun bleached animal bone. Without further inspection we were completely  convinced that some permanent Indian occupational site had existed on Fort Hill.

    Once a site was selected for excavation, it was necessary to assemble the field crew. Reflecting the greater funding available for CWA projects, the initial excavations at the Montague site employed a crew of approximately 30 men (Cadzow 1936:31). Field crews for WPA projects were smaller and more closely reflected pre-FERA archaeological field crew sizes (Setzler 1943:206). Photographs indicate a crew size of approximately 13 men for the Peck site excavations (Figure 4), but crews at later WPA excavations averaged 10 or 11 men, including the project supervisor (Augustine 1938a, 1938c, 1940). While the exact crew size is not known for the Martz Rock Shelters' excavations, existing photographs suggest that a crew of ten was unequally divided between the two sites for this six-day project. Three men apparently excavated Martz Rock Shelter No. 2 (36 SO 223), while the remainder excavated the Martz Rock Shelter (36 SO 14).

    Since the intent of the Somerset County relief excavations was to act as a source of employment for locals, none of the field crew was archaeologists by trade. In a progress report on the 1939 Fort Hill (36 SO 2) excavations, Augustine (1939a:5) noted that:

The excavators are selected from the [Somerset] County's work relief rolls. Some of these men have had previous experience in this work under preceding projects and some have  to be trained. Under this rather unsatisfactory arrangement we have found that the men  are to be commended for their interest in the work, their desire to do their job well and  their help in making the project a success.

    The field supervisor was responsible for maintaining a record of field observations (Augustine 1939a:5). Standard field forms were employed throughout the Somerset County relief excavations to accomplish this task (Figure 5). As with other relief archaeological projects in the United States, field forms were standardized to ensure consistent data recording by unskilled laborers (Dunnell 1986:28). One side of the Somerset form contained places to record provenience information, dimensions of a feature, information on how the feature was located, and associated cultural remains (Figure 5). In some versions of the form, places were also included to record information on burial features. Separate forms were eventually developed and employed for burial features. The front side of the field forms tended to be filled out in more detail and more consistently than the opposite side of the forms. The opposite side of the forms included a grid to record the plan and profile of a feature. Profiles of features were more consistently recorded than their planviews. Available forms have typed information, with some hand-edited corrections, suggesting that they were transferred from original, hand-written forms and notes.

    Equipment for the field excavations came from two principal sources. A magnetic compass, camera, film, field forms, notebooks, and steel tapes were provided by the PHC, which acted as official project sponsor (Augustine 1938f). The field crew had to "supply their own round pointed shovels and four inch pointing trowels" (Augustine 1939a:5-6).

     Once a location had been selected for excavation, and the field crew and necessary equipment had been assembled, an attempt was made to find definitive evidence of a village occupation. Test pits were excavated at potential village locations until features associated with a village occupation were encountered. At least during the 1939 excavations of Fort Hill (36 SO 2), these test pits were 4 x 4 feet square and excavated in five rows (Augustine 1939a). The spacing between these rows of test pits was not recorded.

    After the village occupation was located, the plowzone was stripped, or "topped," from the area of the village occupation using shovels. Parallel stripping trenches ranging from 10 to 20 feet wide were excavated across the village occupation. Trowels were used to remove the last few inches of the top soil above features. This general approach is described in the excavations at the Emerick (36 SO 10) site:

In excavation of occupational sites it is generally customary to disregard the top soil as  deep as plow line, this earth being spaded away. The area between plow line and subsoil  is carefully trowelled with small tools and all discolorations and disturbances are  minutely examined (Augustine 1938b:41).

    While the entire village occupation was excavated, only a portion of the site was open at any one time. Therefore, it was not possible to photograph extensive areas of a village at any one time (Cresson n.d.:17). Photographs from a number of village sites indicate that sticks were placed in postmolds after they were excavated, which enabled aspects of the village community organization to remain visible even after features were backfilled (Figure 6). This was particularly important during winter excavation of sites, where traces of previous field work were quickly covered by snow. Some variation in this general methodology was practiced. For Peck No. 1 (36 SO 1), the stockade was not deliberately traced, but was only distinguished as the stripping progressed (Augustine 1936/7:8). At Fort Hill (36 SO 2), the stockade of the village was encountered during the excavation of stripping trenches 20 feet in width. While most of the crew continued excavating these stripping trenches, two men were tasked with tracing the stockade line (Augustine 1939a:7). Excavation within the village site apparently followed the more standard stripping trench methodology. Three to four foot wide stripping trenches were excavated beyond the boundaries of village sites to ensure that no other village occupations were overlooked. The outer village occupation at Fort Hill (36 SO 2) was found following this methodology (Augustine 1939a:6-7) (Figure 7). Deep test pits were also excavated at each site to ensure that there were no occupational levels below the excavated village site (Cresson n.d.:17).

     The locations of features were mapped from a central station (e.g. site datum) using a Brunton compass to determine bearing; the distance to feature locations was measured using a steel tape (Cresson n.d.:17). This central point was triangulated to relatively permanent aspects of the landscape when possible, including structures and bridges (Augustine 1939a:6). A clear attempt was made to establish a single central station at a site, though in some cases additional central stations were established because of the size of a site or because the initial station was not located close enough to the center of the site (Augustine 1936/7, 1937b). One or two provenience points were recorded for features, depending on the size of the feature (Cresson n.d.:17). Features were assigned a provenience number consisting of the number of the central station and the order the reading was taken. This composite number was placed on field forms and assigned to paper bags containing artifacts collected from features (Augustine 1939a:6).

     Postmold locations were generally only recorded if they were associated with structural features, such as houses, the stockade system, and some post-enclosed features. Not all postmolds associated with a structure were mapped. For houses at most sites, generally six-to-eight locations were shot along the house perimeters and the number of postmolds associated with a house were counted (Cresson n.d.:17). At Fort Hill (36 SO 2), one of the later sites excavated, the average number of recorded postmold locations was higher, ranging from eight-to-twelve locations, and one house had twenty-six post locations recorded (Augustine 1939b). At Peck No. 1 (36 SO 1) and Peck No. 2 (36 SO 8), the number of post-molds associated with houses was generally not counted. Houses at these two sites were recorded with as few as three or four locations along their perimeters, which is an insufficient number to clearly determine house shape independent of Cresson's (n.d.) site maps (Augustine 1936/7, 1937a).

     Features were divided in the field into at least four classes based on either feature morphology or feature contents: fire pits, storage pits, burial pits, and refuse pits. Fire pits contained fire-cracked rock and other evidence of a fire-related function. Post-enclosed features, which were sometimes attached to houses, were classified as storage pits, storage houses, or pit houses (Figure 8). If they were classified as storage pits, the dimensions and attribute information of post-enclosed features were more likely to be recorded. Burial pits obviously contained skeletal remains. Some pits were recorded as possible burial pits even in the absence of skeletal remains, based on largely unspecified aspects of feature morphology and feature contents (see Peck No. 1 (36 SO 1) field form for Burial 3). Refuse pits included all features at a site that did not fit into one of the other three criteria. Butler (1939:20) recognized that so-called refuse pits probably originally had another function and were filled with refuse only after they no longer served that function, a finding which is supported by modern researchers (Dickens 1985; George 1983:18).

     Consistent feature classifications can be assigned to the field data obtained from the Somerset County relief excavations, following procedures similar to those applied by Herbstritt (1981) at Bonnie Brook. Features can be re-classified by shape in planview and profile, as well as by their contents and whether they were enclosed by posts (Figure 9). Metric information on length, width, and depth of feature below plowzone are available for most features; in fact, these were the most consistently recorded feature attributes. It may be possible to assign potential functions to refuse pits based on their dimensions and analogy with features recovered from more recent Monongahela village site excavations.

     Virtually no profile or planview information is available for burial features from sites investigated by the Somerset County relief excavations. The contents of the burial pit were clearly considered more important than the burial pit itself. For burials, the orientation of the burial pit was recorded when possible. The orientation of burial features was indicated by the alignment of the longitudinal axis of the burial feature and the cardinal direction was determined by the end of the burial feature where the head was located. The arrangement of skeletal elements to each other was also described when possible, including the position of the arms and legs relative to the trunk and the direction the head was "facing." If the skeletal remains were too disturbed or poorly preserved to indicate their arrangement, they were described as being "scattered" and that only a "trace" of the skeleton existed. Indications of pit treatment or associated artifacts were noted. In many cases, a clear attempt was made to distinguish intentional burial inclusions from artifacts included incidentally in burial fill. Age and sex of skeletal remains were recorded when possible, though it is not clear in most cases what criteria were used to determine this information.

     With the exception of skeletal elements retained for "museum purposes," human remains encountered in burial features were reinterred after they were photographed and measured in situ (Augustine 1940:53). Due to this practice, only a handful of human skeletal remains from the WPA excavations were available for analysis by T. Dale Stewart, a prominent physical anthropologist affiliated with the Smithsonian Institution. His conclusions were limited, since only one or two partial skeletons from each of five sites were available for study (Stewart n.d.). A modern analysis of the skeletal remains is not possible for most excavated burial features due to the reburial of the remains. Since it was the practice to photograph all burials (Cresson n.d.:18), it is possible to check some burial information against original photographs in many cases.

     A number of problems are inherent in the data from the Somerset County relief excavations, including missing or duplicate feature numbers, missing dimensions, dimensions that do not make sense, transcription errors from field forms, and inconsistent standards for recording feature locations and descriptions. The problems are compounded by the fact that field notes from all the Somerset County relief excavations are incomplete. When discrepancies occur in one part of the data, it is not always possible to check them against original field forms or other field notes. Questionable field procedures were also followed. Winter excavation of some sites, such as Peck No. 1 (36 SO 1), made careful excavation and mapping of features difficult (Figure 10) (Cresson n.d.:18). Mapping problems were compounded by the practice of backfilling features immediately after their excavation (Cresson n.d.:17). Failure to record the location of every postmold at a site makes it difficult to evaluate the accuracy or veracity of structural patterns indicated in site maps, or determine the existence of unrecognized structures.

     Despite the problems found throughout field data collected during the Somerset County relief excavations, it was a relatively straightforward process to transform locational data from these village sites into a format that could be used to evaluate these sites in terms of more modern research questions and for comparison to sites excavated more recently. In most cases, the survey coordinates for features could be converted into more useable grid coordinates. This conversion enables the location of specific feature classes and their contents to be determined in relation to other cultural remains within a village occupation. A more accurate picture of Monongahela community patterns can be developed with this information than is possible from the sketch maps included in Cresson's (n.d.) unpublished manuscript. These sketch maps were evaluated for their accuracy by overlaying the plotted locations of features on a scanned version of the map. Some discrepancies in the field data could be corrected using this procedure. Composite maps can be developed by using Cresson's (n.d.) site plans for bases and by using information from field notes to more clearly associate different feature types with specific locations within sites.
 

• ABSTRACT
• INTRODUCTION
• DEPRESSION-ERA ARCHAEOLOGY IN SOMERSET COUNTY, PENNSYLVANIA
• RECONSTRUCTING SOMERSET COUNTY RELIEF EXCAVATION FIELD METHODOLOGY  <you are here>
• CASE STUDIES, Peck No. 1 (36 SO 1)
•                              Peck No. 2 (36 SO 8)
• CONCLUSIONS
• ACKNOWLEDGEMENTS
• REFERENCES CITED

First appeared in Journal of Middle Atlantic Archaeology 14:39-63.
Original copyright 1998. Reprinted with permission from Archaeological
Services.

Back issues can be ordered at Siftings.com

HOME